Image-based 3D Genomics and Nucleomics in Health and Disease
Date of Award
Spring 2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Genetics
First Advisor
Wang, Siyuan
Abstract
Three-dimensional (3D) chromatin organization influences DNA replication, gene regulation, and cell fate decisions. The last two decades have witnessed an eruption of proximity ligation and sequencing-based methods and image-based methods to study the hierarchical organization and function of 3D chromatin folding in the nucleus. However, it largely remains unexplored how 3D chromatin is hierarchically organized and associated with nuclear landmarks such as the nucleolus and nuclear lamina in single cells, and how such 3D nucleome architectures vary across single cells in situ. It also remains unknown how 3D chromatin organizations are altered in diseases and whether such alterations can be developed into diagnostic and therapeutic biomarkers. In my thesis work, I tried to address these challenges with image-based 3D genomics, nucleomics and transcriptomics techniques. First, I developed a method capable of multiplexed imaging of nucleome architectures (MINA). MINA allowed simultaneous profiling of 3D chromatin folding, RNA expression, and localization of nuclear landmarks in single cells in situ. I identified cell-type-specific 3D chromatin folding associated with gene expression, as well as cell-type-invariant principles of 3D chromatin organizations. MINA revealed multi-faceted features of 3D nucleome architectures in the native tissue context and can be widely applicable to numerous biological contexts in health and disease. Leveraging the spatial transcriptomic profiling capacity of MINA, we investigated into the spatial regulation of hematopoietic stem cells in fetal liver. We identified spatially adjacent pairs of fetal liver cell types and ligand-receptor signaling molecules such as Kitl and Kit. We also identified hematopoietic stem cells, a majority of which are in direct contact with endothelial cells. Second, recent image-based chromatin tracing techniques have identified self-interacting, highly variable single-cell domains that are established and maintained independently of cohesin. Single-cell domain boundaries tend to occur at population-averaged TAD boundaries. To understand epigenetic mechanisms underlying the single-cell domains, we leveraged the active and inactive human X chromosomes with distinct epigenetic landscapes and population-averaged TAD features. We found that the highly variable single-cell domains existed on both active and inactive X chromosomes, even if the inactive X chromosome showed no obvious population-averaged TADs. We also showed that epigenetic perturbations do not disrupt single-cell domain frequencies or strengths. Last but not least, to study how 3D chromatin organization evolved during cancer progression and whether it functionally governed cancer cell states, we performed genome-wide chromatin tracing in a realistic lung adenocarcinoma (LUAD) mouse model. We discovered stereotypical 3D chromatin organizations, including a structural bottleneck at the preinvasive adenoma stage prior to LUAD progression. This indicated a stringent selection on the 3D genome early during cancer progression. We further showed that the 3D genome functionally encodes cancer cell states in single cells, which revealed novel genetic dependencies in LUAD. We also identified a polycomb group protein Rnf2 that partially re-organized the 3D genome through its non-canonical epigenetic activating function. Our results revealed the functional importance of the 3D genome and the potential to identify novel diagnostic and therapeutic biomarkers based on 3D genome organizations. In summary, my thesis work revealed multi-faceted features of nucleome architectures in single cells in situ, explored 3D chromatin folding heterogeneity across single cells, and identified genome-wide chromatin re-organizations that are functionally important during LUAD progression. It demonstrated the power of image-based 3D genomics and nucleomics techniques in generating hypothesis and uncovering new biology in health and disease.
Recommended Citation
Liu, Miao, "Image-based 3D Genomics and Nucleomics in Health and Disease" (2023). Yale Graduate School of Arts and Sciences Dissertations. 971.
https://elischolar.library.yale.edu/gsas_dissertations/971
